APPLICANT'S ABSTRACT: Cognitions regarding drinking and driving have been shown to be predictive of drinking and driving behavior in adolescents. However, no studies have assessed these variables in adolescents who have not begun to drive. As cognitions about a behavior form prior to engaging in that behavior, measuring cognitions of non-driving adolescents can be a crucial step in improving our understanding of how these cognitions develop. The proposed project will explore relationships among a variety of cognitions about drinking, driving, and drinking and driving, behavior. Data will be collected from two samples of adolescents - one of licensed drivers, and one of adolescents who have not obtained a drivers' license. The non-licensed group is hypothesized to have more negative cognitions concerning drinking and driving, and more positive cognitions about alternatives to drinking and driving. A model of the relationships among these cognitive risk variables and their relationship with drinking and driving behavior will be estimated. Finally, mediating and moderating models will also be tested. This research is intended to provide data critical to the theoretical model of the role of cognitions on drinking and driving, and to lay the groundwork for a larger program of research to track changes in these cognitions longitudinally. Improving our understanding of the development of these cognitive risk variables can have important implications for prevention and intervention programs, particularly for programs focused on drivers-in-training.